JP6776707B2 - Own vehicle position estimation device - Google Patents

Description

The present invention relates to a vehicle position estimation device.

Conventionally, Japanese Patent Application Laid-Open No. 2001-331787 is known as a technical document relating to a vehicle position estimation device. In this publication, the first shape candidate of the road on which the own vehicle travels is extracted from the image data captured by the in-vehicle camera, and the second shape candidate of the road is extracted from the map data, and the first shape candidate and the second shape candidate are extracted. A device for estimating the position of the own vehicle on the road based on the overlapping state with the shape candidate is described. This device uses a Kalman filter to collate the overlapping state.

Japanese Unexamined Patent Publication No. 2001-331787

By the way, in order to estimate the position of the own vehicle with high accuracy, a method of further considering the posture (yaw angle, etc.) of the in-vehicle camera is known. However, there is a problem that the yaw angle error accumulates during the curve running of the vehicle, which causes a decrease in the estimation accuracy of the position of the own vehicle.

Therefore, in the present technical field, it is desired to provide an own vehicle position estimation device capable of improving the estimation accuracy of the position of the own vehicle on the map.

In order to solve the above problem, the own vehicle position estimation device according to the present invention is based on the first own vehicle position which is the position on the map of the own vehicle obtained from the in-vehicle positioning unit, and is based on the map information of the own vehicle. It is a vehicle position estimation device that generates a pseudo road surface image of the surroundings and estimates the position of the second vehicle by collating the road surface image obtained from the image captured by the camera of the vehicle with the pseudo road surface image. A collation degree calculation unit that calculates the collation degree with a pseudo road surface image, and a curvature acquisition unit that acquires the curvature of the traveling road on which the own vehicle travels based on the first own vehicle position or the second own vehicle position and map information. A landmark storage unit that stores the position information of the landmarks on the map, and a landmark recognition unit that recognizes the image position coordinates of the landmarks included in the captured image based on the captured image captured by the camera of the own vehicle. A vertical position recognition unit that recognizes the vertical position of the own vehicle in the extending direction of the traveling road based on the position information of the landmark on the map and the image position coordinates of the landmark included in the captured image. The vertical position error recognition unit that recognizes the vertical position error, which is the difference between the vertical position recognized by the vertical position recognition unit and the vertical position of the own vehicle obtained from the second own vehicle position, and the yaw angle of the own vehicle are recognized. The yaw angle recognition unit, the vehicle speed recognition unit that recognizes the vehicle speed of the own vehicle, and the curvature of the traveling road based on the second own vehicle position, the degree of matching, the curvature of the driving road, the vertical position error, the yaw angle, and the vehicle speed. The third vehicle position estimation unit that estimates the position of the third vehicle, which is the position on the map of the vehicle, by the Kalman filter processing set in advance so that the smaller the value is, the larger the observation noise of the yaw rate, which is the time differential of the yaw angle. And.

As described above, according to the own vehicle position estimation device according to the present invention, it is possible to improve the estimation accuracy of the position of the own vehicle on the map.

It is a block diagram which shows the own vehicle position estimation apparatus which concerns on this embodiment. This is an example of a data table for obtaining the value of GE (cc, err_lon) in the formula (5). It is a flowchart which shows the 1st own vehicle position estimation processing. It is a flowchart which shows the own vehicle position estimation process from the second time onward.

Hereinafter, embodiments of the present invention will be described with reference to the drawings.

The own vehicle position estimation device 100 according to the present embodiment shown in FIG. 1 is a device mounted on a vehicle (own vehicle) and estimates the position of the own vehicle on a map. The own vehicle position estimation device 100 estimates the position of the own vehicle with high accuracy by using the collation result of the pseudo road surface image generated from the map information and the road surface image captured by the in-vehicle camera and the Kalman filter.

[Configuration of own vehicle position estimation device]
As shown in FIG. 1, the own vehicle position estimation device 100 according to the present embodiment includes an [Electronic Control Unit] 10 that collectively controls the device. The ECU 10 is an electronic control unit having a CPU [Central Processing Unit], a ROM [Read Only Memory], a RAM [Random Access Memory], a CAN [Controller Area Network] communication circuit, and the like. In the ECU 10, for example, various functions are realized by loading the program stored in the ROM into the RAM and executing the program loaded in the RAM in the CPU. The ECU 10 may be composed of a plurality of electronic control units.

The ECU 10 is connected to a GPS [Global Positioning System] receiver 1, a camera 2, a vehicle speed sensor 3, an IMU [Inertial Measurement Unit] 4, a map database 5, and a landmark database 6.

The GPS receiving unit 1 functions as an in-vehicle positioning unit that measures the position of its own vehicle (for example, the latitude and longitude of the vehicle) by receiving signals from three or more GPS satellites. The GPS receiving unit 1 transmits the measured position information of the own vehicle to the ECU 10. The in-vehicle positioning unit is not limited to the GPS receiving unit 1.

The camera 2 is an imaging device that images the surroundings of the own vehicle. The camera 2 is provided on the back side of the windshield of the vehicle, for example, and images the road surface in front of the vehicle. The camera 2 may be a monocular camera or a stereo camera. The camera 2 transmits the information of the captured image to the ECU 10. The camera 2 may be composed of a plurality of cameras.

The vehicle speed sensor 3 is a detector that detects the speed of the own vehicle. As the vehicle speed sensor 3, a wheel speed sensor that is provided on a wheel of the own vehicle or a drive shaft that rotates integrally with the wheel and detects the rotation speed of the wheel is used. The vehicle speed sensor 3 transmits the detected vehicle speed information to the ECU 10.

The IMU4 is an inertial measurement unit that measures the roll angle, pitch angle, and yaw angle of the own vehicle. The IMU4 is configured to include, for example, a gyro sensor and an acceleration sensor. The IMU4 transmits the detected angle of the own vehicle to the ECU 10. A yaw angle sensor may be used instead of the IMU4. Further, instead of the IMU4, the yaw angle may be obtained from the collation result of the pseudo road surface image generated from the map information and the road surface image captured by the in-vehicle camera.

The map database 5 is a database that stores map information. The map database 5 is formed in, for example, an HDD [Hard Disk Drive] mounted on the own vehicle. Map information includes road position information, road shape information (for example, curve, straight line type, curve curvature, etc.), white line position information, intersection and branch point position information, and building position information. Is done. The map database 5 may be stored in a computer of a facility such as a management center capable of communicating with the own vehicle.

The landmark database 6 is a landmark storage unit provided with landmark information. The landmark database 6 is formed in, for example, an HDD mounted on the own vehicle. A landmark is a fixed position on the road surface (including on a road surface other than the vehicle lane) and serves as a reference for calculating the position of an object. Landmarks include road signs and road markings. Road signs include guide signs, warning signs, regulatory signs, instruction signs, and the like. Road markings include regulatory markings and instruction markings. Regulatory signs include a turn prohibition mark, a maximum speed mark, and the like. The instruction signs include a white line (center line of the road, outside line of the road, lane boundary line, etc.), a diamond mark indicating that there is a pedestrian crossing in front, a triangular mark indicating that there is a priority road ahead, a direction mark, and a traffic light. , Denirator, tunnel entrance / exit, ETC gate entrance / exit, etc.

The landmark database 6 stores the location information of the landmark on the map. That is, the landmark database 6 stores the location information of the landmarks associated with the map information stored in the map database 5. Further, the landmark database 6 stores the image information of the landmark for recognizing the landmark from the image captured by the camera 2. The map database 5 may also serve as the landmark database 6.

Next, the functional configuration of the ECU 10 will be described. The ECU 10 includes a first vehicle position estimation unit 11, a pseudo road surface image generation unit 12, a road surface image conversion unit 13, a collation degree calculation unit 14, a second vehicle position estimation unit 15, and a curvature acquisition unit 16. Further, the ECU 10 has a landmark recognition unit 17, a vertical position recognition unit 18, a vertical position error recognition unit 19, a yaw angle recognition unit 20, a vehicle speed recognition unit 21, and a third own vehicle position estimation unit 22.

The first vehicle position estimation unit 11 estimates the position of the first vehicle, which is the position on the map of the vehicle, based on the positioning result of the vehicle-mounted positioning unit such as the GPS receiver 1 and the map information of the map database 5. To do. The first vehicle position estimation unit 11 estimates the first vehicle position (latitude, longitude) from the positioning results of the vehicle-mounted positioning unit by a well-known method.

The pseudo road surface image generation unit 12 generates a pseudo road surface image around the first vehicle position based on the first vehicle position and map information (position information of the white line). The pseudo road surface image is an image of the road surface of the road simulated from the position information of the white line included in the map information. The pseudo road surface image is generated, for example, so as to correspond to the image of the road surface in a plan view. The pseudo road surface image generation unit 12 generates a pseudo road surface image from the position of the first own vehicle and the map information by a well-known method.

The road surface image conversion unit 13 converts the captured image (image including the road surface) of the camera 2 into a road surface image in a plan view. The road surface image conversion unit 13 converts the captured image of the camera 2 into a road surface image by a well-known method (for example, viewpoint conversion processing) so that the image can be collated with the pseudo road surface image. It is not necessary to convert the road surface image into a plan view if it can be collated with the pseudo road surface image.

The collation degree calculation unit 14 calculates the collation degree between the pseudo road surface image and the road surface image. The degree of collation corresponds to the degree of matching between the pseudo road surface image generated from the map information and the road surface image obtained from the image captured by the camera 2. The higher the degree of collation, the more the pseudo road surface image and the road surface image match. The collation degree calculation unit 14 calculates, for example, the collation degree of the white line of the road as the collation degree between the pseudo road surface image and the road surface image. The collation degree calculation unit 14 calculates the collation degree between the pseudo road surface image and the road surface image by a well-known method.

The second vehicle position estimation unit 15 estimates the position of the second vehicle, which is the position on the map of the vehicle, based on the pseudo road surface image and the road surface image. The second vehicle position estimation unit 15 estimates the second vehicle position (latitude, longitude) from the pseudo road surface image and the road surface image by a well-known method.

The second own vehicle position estimation unit 15 may use the collation degree calculated by the collation degree calculation unit 14. For example, the second vehicle position estimation unit 15 selects a pseudo road surface image having the highest degree of collation among the pseudo road surface images having a degree of collation with the road surface image obtained from the image captured by the in-vehicle camera 2 of a predetermined threshold value or more. Identify. The second vehicle position estimation unit 15 estimates the position of the second vehicle by using the position information on the map on which the specified pseudo road surface image is generated. The second own vehicle position estimation unit 15 estimates the second own vehicle position by collating the pseudo road surface image with the road surface image by a well-known method.

The curvature acquisition unit 16 acquires the curvature of the traveling road on which the own vehicle travels based on the position of the second own vehicle and the map information. The curvature acquisition unit 16 identifies the traveling road on which the own vehicle travels from the position of the second own vehicle, and acquires the curvature of the traveling road based on the information of the road curvature included in the map information. The curvature acquisition unit 16 may acquire the curvature of the traveling road of the own vehicle by using the first own vehicle position instead of the second own vehicle position.

The landmark recognition unit 17 recognizes the landmarks included in the captured image based on the captured image of the camera 2 and the landmark information stored in the landmark database 6. The landmark recognition unit 17 recognizes landmarks in an image by, for example, pattern recognition or edge extraction.

When the landmark recognition unit 17 recognizes the landmark included in the captured image, the landmark recognition unit 17 recognizes the image position coordinates of the landmark in the captured image by well-known image processing. The image position coordinates of the landmark are the coordinates of the position of the image of the landmark in the captured image. For example, the image position coordinates of the diamond mark, which is a road marking on the road surface, can be the coordinates of the center of the diamond mark in the captured image. The landmark recognition unit 17 recognizes the image position coordinates of the landmark by a well-known method.

When the landmark recognition unit 17 recognizes the image position coordinates of the landmark, the vertical position recognition unit 18 of the own vehicle is based on the image position coordinates of the landmark in the captured image and the position information of the landmark on the map. Recognize the vertical position. The vertical position of the own vehicle is the position of the own vehicle in the extending direction of the traveling road on which the own vehicle travels. The vertical position recognition unit 18 recognizes the vertical position of the own vehicle using landmarks by a well-known method.

The vertical position error recognition unit 19 calculates the vertical position error when the vertical position recognition unit 18 recognizes the vertical position of the own vehicle. The vertical position error is the difference between the vertical position of the own vehicle recognized by the vertical position recognition unit 18 and the vertical position of the own vehicle obtained from the second own vehicle position. The vertical position error recognition unit 19 determines the vertical position error by, for example, taking the difference between the second own vehicle position and the vertical position of the own vehicle recognized by the vertical position recognition unit 18 with reference to the extending direction of the traveling road. calculate. The vertical position error recognition unit 19 calculates the vertical position error by a well-known method.

The yaw angle recognition unit 20 recognizes the yaw angle of its own vehicle based on the yaw angle measured by the IMU4. The vehicle speed recognition unit 21 recognizes the vehicle speed of the own vehicle based on the vehicle speed detected by the vehicle speed sensor 3.

When the vertical position error recognition unit 19 recognizes the vertical position error of the own vehicle, the third own vehicle position estimation unit 22 estimates the third own vehicle position, which is the position on the map of the own vehicle. The third vehicle position estimation unit 22 estimates the third vehicle position by preset Kalman filter processing based on the second vehicle position, the degree of collation, the curvature of the road, the vertical position error, the yaw angle, and the vehicle speed. To do. Further, the third own vehicle position estimation unit 22 outputs the yaw angle of the own vehicle in addition to the third own vehicle position as the output result of the Kalman filter processing. Hereinafter, the yaw angle estimated by the Kalman filter processing will be referred to as a second yaw angle.

Hereinafter, the Kalman filter processing according to the present embodiment will be described. The following equation (1) is an equation showing the state quantity u in the Kalman filter processing. In equation (1), E is the latitude of the own vehicle, N is the longitude of the own vehicle, φ is the yaw angle of the own vehicle, V is the vehicle speed of the own vehicle, and φ ^′ (dot in φ) is the yaw rate of the own vehicle. T represents the transpose of the matrix. Equation (2) is an equation showing the observed amount v in the Kalman filter processing. In equation (2), Ec is the latitude at the second vehicle position, Nc is the longitude at the second vehicle position, φc is the yaw angle recognized by the yaw angle recognition unit 20, and Vc is the vehicle speed recognized by the vehicle speed recognition unit 21. φ ^′ c is the yaw rate obtained by differentiating φc.

Next, the observation equation will be described. The observation equation of the Kalman filter processing is shown as the following equation (3). In equation (3), H is an identity matrix. Rn is the observed noise. k represents a time series. In equation (4), equation (3) is represented by a determinant.

The following equation (5) is a determinant showing the observed noise Rn. In the formula (5), cc is the degree of collation calculated by the degree of collation calculation unit 14. err_lon is the vertical position error recognized by the vertical position error recognition unit 19. A ₁ , A ₂ , and A ₃ are constants. _{G E (cc, err_lon),} G R (cc, err_lon), G YR (R) is a function. Equation (6) is an equation showing the function G _YR . In equation (6), R is the curvature of the driving road of the own vehicle. A ₄ , A ₅ , and A ₆ are constants. Equation (6) reflects the characteristic that the observed noise of yaw rate tends to increase when the curvature R is small.

Here, FIG. 2 is an example of a data table for obtaining GE (cc, err_lon) in the equation (5). A shown in FIG. 2 is a constant. The data table of FIG. 2 is set so that the observed noise becomes small when the collation degree cc is large and when the vertical position error err_lon is small. As shown in FIG. 2, if the degree of collation cc and the vertical position error err_lon are determined, the value of GE (cc, err_lon) can be obtained by using a preset data table or the like. G _R (cc, err_lon) for also, it is possible to obtain the value by using the same data table. For example, by replacing the constant a in FIG. 2 to the appropriate constant _b, it can be a data table G R (cc, err_lon).

Next, the equation of state will be described. The equation of state of Kalman filtering is shown as the following equation (7). In equation (7), F is a partially differentiable matrix. Qn is state noise (system noise). k + 1 means that it is the discrete time immediately after the discrete time k. Equations (8) to (12) are shown below as equations expanded from the equation (7) by omitting the state noise Qn. The dt shown in the equations (8) to (10) means the transition time from the discrete time k to the discrete time k + 1.

The following equation (13) is a determinant showing the state noise Qn. In equation (13), B1, B2, B3, and B4 are constants. Fpos (R) is a function of curvature R. Fyaw (φ) is a function of the yaw angle φ. Equation (14) is an equation showing the function Fpos (R). In equation (14), B ₅ , B ₆ , and B ₇ are constants. It is an expression which shows the function Fyaw (R). In equation (15), B ₈ , B ₉ , and B ₁₀ are constants.

The third vehicle position estimation unit 22 can estimate the third vehicle position and the yaw angle by the Kalman filter processing using the equations (1) to (15) described above. According to this Kalman filter processing, by considering the collation degree cc, the vertical position error err_lon, and the curvature R in the observation noise Rn, the estimation accuracy of the third own vehicle position can be improved as compared with the case where these are not taken into consideration. Can be done. Further, by considering the vertical position error err_lon and the curvature R in the state noise Qn, it is possible to improve the estimation accuracy of the third own vehicle position as compared with the case where these are not taken into consideration.

[Own vehicle position estimation process in own vehicle position estimation device]
Hereinafter, the own vehicle position estimation process in the own vehicle position estimation device 100 according to the present embodiment will be described with reference to the drawings. FIG. 3 is a flowchart showing the first vehicle position estimation process. The flowchart of FIG. 3 is implemented, for example, when the own vehicle starts traveling.

As shown in FIG. 3, the ECU 10 of the own vehicle position estimation device 100 estimates the position of the first own vehicle by the first own vehicle position estimation unit 11 as S10. The first vehicle position estimation unit 11 estimates the position of the first vehicle based on the positioning result of the vehicle-mounted positioning unit (for example, GPS receiving unit 1).

In S12, the ECU 10 generates a pseudo road surface image by the pseudo road surface image generation unit 12. The pseudo road surface image generation unit 12 generates a pseudo road surface image around the first vehicle position based on the first vehicle position and map information (position information of the white line). Further, the ECU 10 converts the captured image of the camera 2 into a plan view road surface image (a road surface image that can be compared with a pseudo road surface image) by the road surface image conversion unit 13.

In S14, the ECU 10 calculates the degree of collation between the pseudo road surface image and the road surface image by the collation degree calculation unit 14. The higher the degree of collation, the higher the degree of matching between the pseudo road surface image and the road surface image.

In S16, the ECU 10 estimates the position of the second vehicle by the second vehicle position estimation unit 15. The second vehicle position estimation unit 15 estimates the position of the second vehicle based on the pseudo road surface image and the road surface image. The second vehicle position estimation unit 15 obtains, for example, a pseudo road surface image having the highest degree of matching among pseudo road surface images having a degree of matching with the road surface image obtained from the image captured by the in-vehicle camera 2 of a predetermined threshold value or more. The position of the second own vehicle is estimated by using the position information of the map on which the specified pseudo road surface image is generated.

In S18, the ECU 10 acquires the curvature of the traveling road on which the own vehicle travels by the curvature acquisition unit 16. The curvature acquisition unit 16 identifies the traveling road on which the own vehicle travels from the second own vehicle position based on the second own vehicle position and the map information, and acquires the curvature of the traveling road from the map information.

In S20, the ECU 10 determines whether or not the landmark recognition unit 17 has recognized the landmark included in the image captured by the camera 2. The landmark recognition unit 17 recognizes the landmarks included in the captured image based on the captured image of the camera 2 and the landmark information stored in the landmark database 6. When the ECU 10 determines that the landmark included in the captured image has been recognized (S20: YES), the process proceeds to S22.

On the other hand, when the ECU 10 determines that the landmark included in the captured image is not recognized (the landmark does not exist in the captured image) (S20: NO), the process proceeds to S28.

In S22, the ECU 10 recognizes the image position coordinates of the landmark in the captured image by the landmark recognition unit 17. The landmark recognition unit 17 can recognize the coordinates of the center of the rhombus mark in the captured image as, for example, the image position coordinates of the rhombus mark which is a road marking.

In S24, the ECU 10 recognizes the vertical position of the own vehicle by the vertical position recognition unit 18. The vertical position recognition unit 18 recognizes the vertical position of the own vehicle based on the image position coordinates of the landmark in the captured image and the position information of the landmark on the map.

In S26, the ECU 10 calculates the vertical position error by the vertical position error recognition unit 19. The vertical position error recognition unit 19 calculates the vertical position error as the difference between the vertical position of the own vehicle recognized by the vertical position recognition unit 18 and the vertical position of the own vehicle obtained from the second own vehicle position.

In S28, the ECU 10 estimates the third vehicle position and the second yaw angle by the third vehicle position estimation unit 22. The third vehicle position estimation unit 22 uses a preset Kalman filter process based on the second vehicle position, the degree of matching, the curvature of the road, the vertical position error, the yaw angle, and the vehicle speed, and the third vehicle position and the third vehicle position. 2 Estimate the yaw angle. The third vehicle position estimation unit 22 can estimate the third vehicle position and the second yaw angle by, for example, Kalman filter processing using the above equations (1) to (15). The ECU 10 outputs the estimated third vehicle position as a position on the map of the vehicle to the outside (for example, a navigation system, a driving support device, an automatic driving device). Here, in the case of NO in S20 (when the landmark is not recognized), the Kalman filter processing is executed assuming that the vertical position error is the maximum value set in advance. In this case, in the table of FIG. 2, the value in the column of "2 ≦ err_lon", which is the maximum value, is used in relation to the degree of collation cc.

In addition to the flowchart of FIG. 3, the yaw angle recognition unit 20 recognizes the yaw angle of the own vehicle and the vehicle speed recognition unit 21 recognizes the vehicle speed of the own vehicle at predetermined intervals according to the performance of the sensor. There is.

FIG. 4 is a flowchart showing the own vehicle position estimation process from the second time onward. The flowchart shown in FIG. 4 is executed, for example, when the third own vehicle position and yaw angle are obtained by the first Kalman filter processing after the own vehicle starts traveling.

As shown in FIG. 4, the ECU 10 generates a pseudo road surface image by the pseudo road surface image generation unit 12 as S30. When the position of the third vehicle is estimated once, the pseudo road surface image generation unit 12 is around the position of the third vehicle based on the position of the third vehicle and the map information (position information of the white line) with high estimation accuracy. Generate a pseudo road surface image. Further, the ECU 10 converts the captured image of the camera 2 into a plan view road surface image (a road surface image that can be compared with a pseudo road surface image) by the road surface image conversion unit 13.

In S32, the ECU 10 calculates the degree of collation between the pseudo road surface image and the road surface image by the collation degree calculation unit 14. This pseudo road surface image is a pseudo road surface image around the position of the third vehicle.

In S34, the ECU 10 estimates the position of the second vehicle by the second vehicle position estimation unit 15. The second vehicle position estimation unit 15 estimates the position of the second vehicle this time based on the pseudo road surface image and the road surface image around the third vehicle position.

In S36, the ECU 10 acquires the curvature of the traveling road on which the own vehicle travels by the curvature acquisition unit 16. The curvature acquisition unit 16 acquires the curvature of the traveling road based on the position of the second vehicle and the map information.

In S38, the ECU 10 determines whether or not the landmark recognition unit 17 has recognized the landmark included in the image captured by the camera 2. When the ECU 10 determines that the landmark included in the captured image has been recognized (S38: YES), the process proceeds to S40.

On the other hand, when the ECU 10 determines that the landmark included in the captured image is not recognized (S38: NO), the process proceeds to S46.

Since S40 to S44 are the same processes as S22 to S26 shown in FIG. 3, the description will be simplified. In S40, the ECU 10 recognizes the image position coordinates of the landmark in the captured image by the landmark recognition unit 17. In S42, the ECU 10 recognizes the vertical position of the own vehicle by the vertical position recognition unit 18. In S44, the ECU 10 calculates the vertical position error by the vertical position error recognition unit 19.

In S46, the ECU 10 estimates the current third vehicle position and the current second yaw angle by the third vehicle position estimation unit 22. The third vehicle position estimation unit 22 uses a preset Kalman filter process based on the second vehicle position, the degree of collation, the curvature of the driving road, the vertical position error, the previous second yaw angle, and the vehicle speed. 3 Estimate the position of the vehicle and the second yaw angle this time.

The third vehicle position estimation unit 22 can estimate the current third vehicle position and the current second yaw angle by, for example, Kalman filter processing using the above equations (1) to (15). .. In this case, the previous second yaw angle (yaw angle estimated by the previous Kalman filter processing) is used as the yaw angle φ in the equation (1).

It is not always necessary to use the previous second yaw angle, and the Kalman filter process may be executed using the yaw angle of the own vehicle recognized by the yaw angle recognition unit 20. Further, in the case of NO in S38 (when the landmark is not recognized), the Kalman filter processing is performed by utilizing the vertical position error used in the previous processing. The ECU 10 outputs the estimated position of the third own vehicle to the outside as a position on the map of the own vehicle. Then, after the elapse of a predetermined time, the ECU 10 repeats the process from S30 again using the position of the third own vehicle this time.

[Effect of own vehicle position estimation device]
According to the own vehicle position estimation device 100 according to the present embodiment described above, the second own vehicle position, yaw angle, and vehicle speed estimated by collating the pseudo road surface image generated from the map information with the road surface image captured by the in-vehicle camera can be obtained. By performing Kalman filter processing that takes into account the curvature of the driving road and vertical position error as input, it is possible to correct the yaw angle based on the road curvature as compared to the case where the road curvature and vertical position error are not taken into consideration. It is possible to improve the estimation accuracy of the position on the map.

Further, according to the own vehicle position estimation device 100, as shown in FIG. 4, after estimating the third own vehicle position once, the next own vehicle position estimation process is performed using the third own vehicle position. Therefore, it is not necessary to use an in-vehicle positioning unit such as the GPS receiving unit 1 every time, and high estimation accuracy can be maintained for the position of the own vehicle on the map.

Although the preferred embodiment of the present invention has been described above, the present invention is not limited to the above-described embodiment. The present invention can be carried out in various forms having various modifications and improvements based on the knowledge of those skilled in the art, including the above-described embodiment.

For example, the second vehicle position estimation unit 15 does not need to use the collation degree in estimating the second vehicle position. In this case, the collation degree calculation unit 14 may calculate the collation degree between the pseudo road surface image and the road surface image after estimating the position of the second own vehicle. That is, S14 in FIG. 3 may be performed after S16. Further, the acquisition of the curvatures of the traveling roads of S10 to S16 and S18 may be completed before the Kalman filter processing of S28, and the order is not limited. Similarly, S32 in FIG. 4 may be performed after S34. The acquisition of the curvatures of the roads of S30 to S34 and S36 may be completed before the Kalman filter processing of S46, and the order is not limited. In the flowchart showing the first vehicle position estimation process shown in FIG. 3, if NO in S20 (when the landmark is not recognized), the Kalman filter process in S28 is not performed, and only in the case of YES in S20, S28. The Kalman filter process may be executed.

100 ... Vehicle position estimation device, 10 ... ECU, 1 ... GPS receiver (positioning unit), 2 ... Camera, 3 ... Vehicle speed sensor, 4 ... IMU, 5 ... Map database, 6 ... Landmark database (landmark storage unit) ), 11 ... 1st vehicle position estimation unit, 12 ... pseudo road surface image generation unit, 13 ... road surface image conversion unit, 14 ... collation degree calculation unit, 15 ... second vehicle position estimation unit, 16 ... curvature acquisition unit, 17 ... Landmark recognition unit, 18 ... Vertical position recognition unit, 19 ... Vertical position error recognition unit, 20 ... Yaw angle recognition unit, 21 ... Vehicle speed recognition unit, 22 ... Third vehicle position estimation unit.

Claims (1)

Based on the position of the first own vehicle, which is the position on the map of the own vehicle obtained from the in-vehicle positioning unit, a pseudo road surface image around the own vehicle is generated from the map information, and the captured image of the camera of the own vehicle is generated. This is a vehicle position estimation device that estimates the position of the second vehicle by collating the road surface image obtained from the above with the pseudo road surface image.
A collation degree calculation unit that calculates a collation degree between the road surface image and the pseudo road surface image,
A curvature acquisition unit that acquires the curvature of the traveling road on which the own vehicle travels based on the first own vehicle position or the second own vehicle position and the map information.
A landmark storage unit that stores the location information of landmarks on the map,
A landmark recognition unit that recognizes the image position coordinates of the landmark included in the captured image based on the captured image captured by the camera of the own vehicle.
Based on the position information of the landmark on the map and the image position coordinates of the landmark included in the captured image, the vertical position which is the position of the own vehicle in the extending direction of the traveling road is recognized. Vertical position recognition unit and
A vertical position error recognition unit that recognizes a vertical position error that is a difference between the vertical position recognized by the vertical position recognition unit and the vertical position of the own vehicle obtained from the second own vehicle position.
The yaw angle recognition unit that recognizes the yaw angle of the own vehicle and
The vehicle speed recognition unit that recognizes the vehicle speed of the own vehicle and
Based on the second vehicle position, the degree of collation, the curvature of the traveling road, the vertical position error, the yaw angle, and the vehicle speed, the smaller the curvature of the traveling road, the more the yaw rate is the time derivative of the yaw angle. A third vehicle position estimation unit that estimates the position of the third vehicle, which is the position on the map of the vehicle, by Kalman filter processing preset so as to increase the observation noise .
A vehicle position estimation device equipped with.

Images